plantimals

this a sign that you should dump the context and start on the feature from the specs up. context engineering is real https://youtu.be/IS_y40zY-hc nostr:nevent1qqsvhgjv74f0kk6elsrte7hcjq48kly8gr6kfm8wfdqyzcay9234q8cpzamhxue69uhhyetvv9ujumn0wd68ytnzv9hxgtczyqt0rgqsp4x0l77vggcw3c8y9yxvtpyurtwxf4n98ldq0sp3kyr5kqcyqqqqqqgpxr9gm

self-improvising memory. it rhymes with AI https://www.mdpi.com/1099-4300/26/6/481

context engineering seems like a valid concept. I've had good results following this approach https://github.com/humanlayer/advanced-context-engineering-for-coding-agents/blob/main/ace-fca.md

from Erik Hoel: https://drss.io/DNAl

finally got my claude code set up the way I like it https://www.youtube.com/watch?v=VddS5IWxHd8

great podcast episode from nostr:nprofile1qy2hwumn8ghj7etyv4hzumn0wd68ytnvv9hxgqgewaehxw309a4kjarrdpjkutn6v9czucm0da4kjmn8qqsry0gl27g2q54ae44lq4kh3st9dyjy7mlsz865q7hw3nm097nsles62z04z Luke Gromen "Print the Money or Trigger the Revolution" demonetization of farmland by #bitcoin nostr:naddr1qvzqqqr4gupzppk0xapy8t0jc5txkguhugc26lgapzkcuvkqqh3vv056nfncuwhgqqm5cattv5k5wun0d4jkutfzwpexjmn5946xsefdd4hkuete94hhytt5wf5kwem9wgkhg6r994ex2an0d36hg6t0dc3qkt249z

most of the insights of The Wealth of Networks have since diffused into our awareness. it sees mostly positive feedbacks and not much negative, but still worth a look for those working on snd thinking about the protocol. https://cyber.harvard.edu/wealth_of_networks/Main_Page

testing drss link shortener: https://drss.io/w9p7

a funny aspect of getting the coding agent to write meaningful tests around #nostr functionality is the understanding that secret key -> pubkey cannot be faked. I see so many "npub123-abc" in its tests. no matter how often I explicitly instruct it otherwise. I have a rule in CLAUDE.md which points out that we have a keys.md with a markdown table with 10 secret keys, pubkeys, and npubs for use in tests, and it always forgets and just makes things up.

twitter is the uber of microblogging

I keep coming back to making lists of lists, or more accurately, set of sets. is there a reason why it doesn't exist in the NIPS? just hasn't been a need so far, or was it discussed and discarded? I ask because I've looked and did not find anything. thinking of using 30100 for that kind. I like the idea of a hierarchical list of lists than can link in lists owned and maintained by others. if you hang follow sets from those lists, it becomes a nice way to subscribe to a feed, if there are clients that allow it.

this is a common issue I think, but specifically with claude code.. I'm so exhausted by having to explain when I'm asking a question about a proposed change I am never asserting something... always just asking. when i ask something like "why are you using Getenv() here?" I usualy get: "You're absolutely right! I was so wrong, I'll stop that and go this other way..." if I wanted to express that sentiment I'd do it directly. has anyone found a way to phrase a question economically that conveys that you actually just want an answer?

This Week in Bitcoin from nostr:nprofile1qqswhzptpwm9n0mjydgzpg9csnz20kqhuzhnjqm32umtz3sc3vwss6qpzpmhxue69uhkumewwd68ytnrwghszxthwden5te0wfjkccte9eekummjwsh8xmmrd9skctcpz4mhxue69uhhyetvv9ujuerpd46hxtnfduhs86vymj https://npub.blog/a/naddr1qvzqqqr4gupzp4hx023yj4tkmanpmnhf50qu4vl78ukwze895w9c7ckxj9vfauz8qqgrwv36942xsefd2dcxzmfd2ashy6896ln/ npub.blog translates articles with imeta tags containing 'audio/mpeg' mimetypes into podcast episodes

I am using claude-code, and one constant struggle I have is getting it answer questions without poisoning it's concept of what we're working on. I create markdown files describing the features and such, but when I come across strange implementation and ask something like: "is there an implementation of this in go-nostr we could use instead of writing new code?" it takes that as a suggestion "You're absolutely right!" and runs off to edit the code, rather than just answering my question. I have various lines in my CLAUDE.md describing how not to run off and change things without being explicitly told to do so, but that all eventually falls out of context or is just not working. do any of you have suggestions on how to ask claude code questions in context without derailing things? #ai